SysAdmin - Essential commands

In

Linux Foundation Certified SysAdmin (LFCS): Essential commands

Log into local and remote consoles

  • Connect via SSH (text mode)

    1
    2
    3
    4
    5
    # no userName takes the one you are logged with on your machine
    # machine can be an IP address or a domain
    ssh username@machine
    # ECDSA key is added the first time
    # the  insert password

  • Connect via VNC (grahical mode - VNC viewer)

    • Give the network address
    • Check ECDSA key (fingerprint) to connect
    • Add username and password

Warning!
The server we want to connect to must have xwindows installed, plus VNC server

Search for files

Linux is case sensitive

  • Find

    • Basic
      1
      2
      3
      4
      5
      6
      7
      touch etc/test.txt
      # find on current directory
      find -name "test.txt"
      # check what is on our current directory
      ls -la
      # find starting from root
      find / -name "test.txt"
    • Advanced
      • by name
        1
        2
        3
        4
        5
        6
        # find starting from etc
        find /etc -name "test.txt"
        # -i to ignore case sensitive
        find /etc -iname "test.txt"
        # reverse search
        find /etc -not -name "test.txt"
      • by type
        1
        2
        3
        4
        5
        6
        7
        8
        # find for type (e.g. character devices)
        find /etc -type c
        # find symbolic links
        find /etc -type l
        # find symbolic directories
        find /etc -type d
        # find files whose name ends in ''.log'
        find /etc -type -f "*.log"
      • by size
        1
        2
        # files whose size is greater than 27K
        find /usr/bin -size +27000
      • by creation date
        1
        2
        3
        4
        # created more than 1 day ago
        find /usr/bin -type f -mtime 1
        # created less than 1 day ago
        find /usr/bin -type f -mtime -1
      • by user creation date
        1
        2
        # owned by myUser, pipe it through more to pause on each screen
        find /usr/bin -user myUser | more
      • by permision
        1
        2
        3
        4
        5
        # with permissions 755
        find /usr/bin -perm 755
        # you may do things which what you find on the same line
        # '{} \;' are placeholders
        find /usr/bin -name "test.txt" -exec chmod 700 {} \;

  • Check petitions

    1
    2
    3
    4
    # df stands for disk-free
    # - h stands for human readable
    df -h
    # root, mounted devices, temporary file systems...

  • Find executable in our path

    1
    2
    which python
    which ifconfig

  • Manual

    1
    man find

  • Locate

    1
    2
    3
    # Less powerful, depends on database which should be updated
    updatedb
    locate test.txt

Basic file system features and options

  • Definitions
    • block device: set of addressable blocks used to store and retrieve data
    • filesystem: where a computer system persists general data for users and/or applications

Filesystems can affect

  • Performance pf the system
  • Efficiency of the media
  • Compatibility with other systems
  • Jounaling
    • journal prevent data corruption, records metadata
    • add a bit of overhead to write files
    • some high performance servers might not need it
    • often not used on removable media
  • Current standard
    • EXT
      • EXT (EXTended File System)
      • EXT2 (extended file attributes)
      • EXT3 (journaling)
      • EXT4(add backward compatibility)
  • Coming standard
    • BtrFS (substitute to EXT by Oracle)
      • B-tree File system
      • drive pooling, snapshots, compression, online defragmentation
  • Common alternatives
    • ReiserFS
      • New features that wouldn’t be implemented by EXT (efficiencies for small text files)
      • Unlikely to continue development
    • ZFS (Sun, for Solaris, acquired by Oracle - opensourced not compatible with Linux kernel license)
      • drive pooling, snapshits, dynamic disk striping
      • each file has a checksum
      • (Ubuntu 16.04 supports it) uses containers by default
    • XFS
      • similar to Ext4
      • can be enlarged (but not shrunk) on the fly (good for backup servers)
      • bad for many small files (like web servers)
    • JFS (Journaled, by IBM)
      • low CPU usage
      • partitions can be enlarged (but not shrunk)
      • support in most every major distribution
      • not widely tested
  • Crossplatform
    • FAT (FAT16, FAT32, exFAT)
      • Microsoft File Allocation Table, not journaled
      • USB drives on Windows and Apple hardware
  • Swap space
    • virtual memory ()scratch space)
    • hibernating (analogous to Windows Paging File)

Compare and manipulate file content and I/O redirection

  • List all files
    1
    2
    # list files
    ls -la
  • Read (send to default output) and concatenate
    1
    2
    3
    4
    5
    6
    7
    8
    # catenate
    cat shopingListfile.txt
    # catenate various files as a single file
    cat shopingListfile.txt anotherList.txt
    # catenate, pipe it to paginate it with 'more'
    cat shopingListfile.txt | more
    # pip it to navegate using the up and down arrow keys
    cat shopingListfile.txt | less
  • Sort
    1
    2
    3
    4
    # sort the text content in alphabetical order
    sort shopingListfile.txt | more
    # sort the text content in reverse alphabetical order
    sort -r shopingListfile.txt | more
  • Create a new output file with >
    1
    cat shopingListfile.txt anotherList.txt | sort > combined.txt
  • Format (check man)
    1
    2
    # fix spacing issues on a file
    fmt -u format.txt
  • Numbered lines
    1
    2
    # If it si not redirected to other file, the numbers won't be stored
    nl format.txt > numberedFormat.txt
  • Remove delimiters
    1
    2
    3
    # -d means delimited
    # -f1 means first part
    cut -d ";" -f1 delimited.txt
  • Redirection
    1
    2
    3
    4
    # store on a new file, or overwrite if it already exists
    ls -la > directory.txt
    # append to that file
    ls -ls >> directory.txt

Basic Regular Expressions

  • Anchor characters (^, $)

    1
    2
    3
    4
    # starts the line with A
    ^A
    # ends the line with Z
    Z$

  • Selectors

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    # Find all the lines that start with 'The'
    grep '^The' alice.txt
    # Find line with a single character - '.' means any character
    grep '^.$' alice.txt
    # Find line which starts with a character
    grep '^[0123456789]' alice.txt
    # Find line which starts with a character, short version
    grep '^[0-9]' alice.txt
    # With '*' A character appear 0 or more times (e.g at least has 2 digits)
    grep '^[0-9]*[0-9][0-9]' alice.txt
    # Between 4 and 8 lowercase characters
    grep '[a-z]\{4,8\}' alice.txt
    # Search for a word with either capital or lowercase letter
    grep '\<[tT]he>' alice.text
    # back-reference: any 2 lowerkey letters adjoined (e.g. aa, baac)
    # [a-z][a-z] won't work for that pairs
    grep '\([a-z]\)\1' alice.txt

Archive, backup, compress, unpack, and decompress files

  • archive and backup (tar)
    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    cd /home/user
    #backup on dir
    # create dir
    mkdir datab_kup
    cd data_bakup
    # recursive copy of the '/data' folder
    cp -r /data .
    # undo hat was done previously
    rm *
    rm -rf *
    # backup the 'data' folder with tar (from 'tape archives')
    # concatenate, no compression
    # -c create new file
    # -v for verbose
    # -f for file
    tar cvf databkup.tar data
    # -t lists content of tar file
    tar tvf databkup.tar
    # search files on tar for a particular the text `wpa_supplicant`
    tar tvf databkup.tar | grep wpa_supplicant
  • compress (zip)
    1
    gzip databkup.tar
  • archive plus compress
    1
    2
    # add the 'z' directive
    tar cvfz databkup.tar.gz /data
  • extract or unzip/decompress
    1
    2
    # extract everything on a new directory
    tar xvfz databkup.tar.gz

Create, delete, copy, and move files and dirs

  • create
    1
    2
    3
    4
    5
    6
    7
    8
    9
    # file
    touch testFile
    # edit the file with pico or nano
    nano testFile
    # directory
    mkdir secondFiles
    touch testFileDir
    # create several directories (one inside the other)
    mk -p secondFiles/textFiles/otherFiles
  • remove
    1
    2
    3
    4
    5
    6
    # remove file
    rm testFile
    # remove directory
    rmdir secondFiles
    #remove directory tree, recursive, force
    rm -rf textFiles
  • copy 
    1
    2
    # copy to testfile2
    cp testFile testFile2
  • move and rename
    1
    2
    3
    4
    # move file
    mv testfile2 secondFiles/
    # rename file via mv
    mv testFile2 testFile
  • Hard link: links the original data chunk on disk.Hard links cannot traverse filesystems (physical or shared)
    1
    2
    3
    # create a hard link to a file
    ln ../dir2/file2 hardLinkToFile2
    # deleting the original link doesn't lose the data
  • Soft (symbolic) link: links the original link 
    1
    2
    3
    4
    # create a soft link to a file
    ln -s ../dir2/file2 softLinkToFile2
    # on 'ls' it is represented with an arrow ->
    # deleting the original link loses the data, as a direct access

List, Set, and change standard file permissions

  • Get the list

    1
    2
    # list files
    ls -la

  • Output

    1
    2
    3
    4
    5
    total 8
    drwxr-xr-x  2 user1  user    4096  Feb 28 14:26  .
    -rw-r--r--  1 user1  dbadmin  0     Feb 28 14:26  user1andgroup
    -rw-r--r--  1 user1  user1    0     Feb 28 14:26  user1only
    -rw-r--r--  1 user1  user1    0     Feb 28 14:26  user1yone

    Sets meaning:

    1. directory (‘d’ or ‘-‘’)
    2. ownerPermission-groupPermission-otherPeoplePermission (‘r’, ‘w’, ‘x’ or ‘-‘)
    3. Owner

  • Change permissions

    1
    2
    3
    4
    5
    6
    7
    8
    # grant permissions for user chad, to grant
    chmod u+rwx user1only
    # take away group permissions
    chmod g-rwx user1only
    # take away other users permissions
    chmod o-rwx user1only
    # grant permissions to everyone else
    chmod a+rwx user1only

Permissions via octal numbers

  • 4: read
  • 2: write
  • 1: exec
    You may do all 3 commands in a single one with the octal numbers,
    so to grant all permissions for user, group and other would be 4+2+1=7 -> chmod 777 user1Only
  • Change ownership of files 
    1
    2
    3
    4
    5
    # find files from lostUser
    sudo find / -user lostUser
    # supose the results says they are on '/opt/myapp'
    # change ownership to cloud_user from files on a certain folder
    sudo chown -R cloud_user:cloud_user /opt/myapp

Read and use system documentation

  • Manual
    1
    2
    3
    4
    # manual + command we want to know about, q to exit
    man ls
    # update the man db
    mandb
  • apropos
    1
    2
    # commands related to what you asked about
    apropos ps
  • info
    1
    2
    # similar to man
    info ipc
  • locate
    1
    2
    # its db is updated by a cronjob once a day
    info systemctl.conf
  • the documentation folder: user/share/doc

Manage access to the Root account

  • Get superuser permission
    1
    sudo ls
  • Grant sudo permission
    1
    2
    3
    4
    sudo visudo
    # use either vim or nano
    # find part to grant and add
    # chad ALL=(ALL)  ALL
  • Be added to the WHEEL (CentOS) or SUDO (Ubuntu) group
    1
    2
    3
    4
    5
    6
    7
    8
    # show groups for user1
    groups user1
    # add user1 to group WHEEL
    # -a is important: it adds, if not, it will switch group
    usermod -a -G wheel user1
    # if it fails, do "sudo bang-bang"
    sudo!!
    sudo usermod -a -G wheel user1

Securely copy a file to another server

1
2
# scp file_name userId@destinyServer:destinyFolder
scp myapi.tar.gz cloud_user@SERVER_TWO_PUBLIC_IP:/home/cloud_user/myapi.tar.gz